Measuring microamps & milliamps at 3 MHz bandwidth, AD8428

carl0s

carl0s

Member
Hi
I came across this article written by Paul. He uses an AD8248 as an amplifier for very low voltage changes across a shunt resistor, to "see" the current on the scope, at high bandwidth.

I was wondering, since it was written more than 8 years ago, would this still be a good way to do something similar? I probably don't need the 3MHz bandwidth, but I would like to use my scope to see very small current changes on a microcontroller, and why not make it better if possible. Also I am reading (soon.. when I get chance) Colin O'Flynn's hardware hacking handbook which is all about watching current draw while certain processor ops or code loops take place, so it could be useful there. Perhaps if I read the book it would tell me how to do this... I have been skimming through it, but backwards, dipping in the middle etc, which is stupid..

Is there is now a better or more suitable, or maybe mostly as good but cheaper (this chip is now $30 USD or more in the UK) option now, 8+ years later?

Also, there isn't an explanation in the article why a 5V to +/- 15V power supply was needed. Does the gain amount correspond to the input voltage on the AD8248 or something? Sorry if that's a stupid question, I'm a beginner.

Thanks,
Carl
 
If you use a shunt on the ground connection this ought to be doable without amplification in many cases.

If you monitor the voltage across a shunt diode you can see currents over many orders of magnitude, although the calibration
won't be so good.

One issue with any shunt measuring system is that the shunt either goes between the chip and its decoupling (difficult with
modern chips) or you lose bandwidth due the decoupling caps limiting the bandwidth seen at the shunt.

The AD8248 isn't rail-to-rail so it needs a power supply with headroom above and below the input voltages.

For high accuracy at low bandwidth an auto-zeroing opamp could be used, these can have input errors in the sub-microvolt
range.

In practice start with a good 'scope mode to monitor a ground-side shunt resistor. You'll then be in a position
to judge if you need more bandwidth or voltage sensitivity. You can try probe at x10 and x1 for comparison (trading
sensitivity for bandwidth).
 
carl0s:
Before you progress further I would suggest you peruse the Web page for this IC (Instrumentation Amplifier) at https://www.analog.com/en/products/ad8428.html#product-overview. On that page there is a button to do a search for other Instrumentation Amps there may have similar characteristics.
If you could further tell us what you are planning to accomplish with specific details then perhaps we could be of more help. Otherwise anyone who might seek to assist you would only be shooting in the dark.
Remember the general forum rule at the top of each page in this forum. (THE DEVIL IS IN THE DETAILS!).

Regards,
Ed
 
The voltage gain you need is pretty much fixed, since you have a certain shunt resistor in mind and need a few volts for a normal scope to measure well.

So the main question is how much bandwidth is needed.

If you want several MHz to see rapid changes and high voltage gain, then you're limited to pretty special & expensive amplifiers. High gain-bandwidth costs money.

If you don't need high bandwidth, many really nice parts exist. One of my recent favorites is INA190, because it's so convenient and has low offset error. But you only get time response in the dozens of microseconds range.
 
Thank you Paul. That part looks interesting. The 6-in-1 eval kits look very interesting.
cheers :)
Carl
 
MarkT said:
If you use a shunt on the ground connection this ought to be doable without amplification in many cases.

If you monitor the voltage across a shunt diode you can see currents over many orders of magnitude, although the calibration
won't be so good.

One issue with any shunt measuring system is that the shunt either goes between the chip and its decoupling (difficult with
modern chips) or you lose bandwidth due the decoupling caps limiting the bandwidth seen at the shunt.

The AD8248 isn't rail-to-rail so it needs a power supply with headroom above and below the input voltages.

For high accuracy at low bandwidth an auto-zeroing opamp could be used, these can have input errors in the sub-microvolt
range.

In practice start with a good 'scope mode to monitor a ground-side shunt resistor. You'll then be in a position
to judge if you need more bandwidth or voltage sensitivity. You can try probe at x10 and x1 for comparison (trading
sensitivity for bandwidth).
Click to expand...

Thanks Mark!
 
sbfreddie said:
carl0s:
Before you progress further I would suggest you peruse the Web page for this IC (Instrumentation Amplifier) at https://www.analog.com/en/products/ad8428.html#product-overview. On that page there is a button to do a search for other Instrumentation Amps there may have similar characteristics.
If you could further tell us what you are planning to accomplish with specific details then perhaps we could be of more help. Otherwise anyone who might seek to assist you would only be shooting in the dark.
Remember the general forum rule at the top of each page in this forum. (THE DEVIL IS IN THE DETAILS!).

Regards,
Ed
Click to expand...

Thanks Ed. I need to catch up on work & sleep. I'll be coming back to my project in a week or two :)
 
You must log in or register to reply here.
Back
Top